Every small business that decides to “professionalize its management” follows the same impulse: buy the most complete system the budget allows. An ERP with an inventory module, a CRM with pipeline automation, a management platform with ten different screens. The logic seems obvious — more capability, more control. In practice, it’s the opposite: the more complete the system, the higher the chance it becomes an expensive version of the messy spreadsheet it was supposed to replace.
The problem isn’t the system. It’s the question the company forgot to ask before buying it.
The system doesn’t create the process, it only speeds up what already exists
An eight-person accounting office decides to bring in an ERP to organize client service. The system is good — it has a field for everything, a report for everything, automation for everything. Six months later, half the team still handles client questions over personal WhatsApp, because there was never a formal process for “how a client request comes in, gets triaged, and gets answered” — it only ever existed in the heads of the two most senior people on the team. The system had nothing to run on top of, because there was no process there to begin with. It became an expensive contact database nobody uses the way it was meant to be used.
This isn’t a training problem, and it isn’t “the team resisted change.” It’s a sequencing error: the company tried to automate a process that had never been designed in the first place. No system fixes that, because a system doesn’t create process — it only executes faster what the company already knows how to do. If what the company knows how to do lives only in someone’s head, the most expensive system on the market changes nothing.
It’s the same logic João Paulo Batistella, an innovation executive with a career spanning Ericsson, Telefônica/Vivo, and five years as CEO of EISA, applies to AI adoption in companies: formalized data and process come before automation, not after. A routine that exists only as tacit knowledge has no foundation for any tool — AI or otherwise — to act on. The same reasoning applies, without exception, to any management system a small business considers buying.
The wrong question is “which system to buy”
The right question isn’t which tool has the most features. It’s: what business decision am I unable to make today, and would this system solve that? A system isn’t a work tool, it’s the answer to a specific decision question. If the question isn’t clear, no system — however robust — will make a difference, because the company doesn’t know what to ask it to do.
A limited budget isn’t an excuse to think small — it’s a reason to think with more discipline. A small business doesn’t have the margin to buy a complete system and discover, a year later, that it only uses 10% of it. It needs to get the priority right on the first attempt.
Framework: 3 questions before any new system
1. Is this process already formalized, or would the system be the first time anyone designs it? If the answer is “I can only explain it out loud,” the problem to solve first isn’t technological — it’s documenting the rule, with an explicit decision about what’s in and what’s out. Buying a system before that just means paying a lot to digitize the mess.
2. What business decision gets better with that organized data? Not “what task gets faster” — a faster task without a better decision is just a bit less operational cost, which is fine, but a marginal gain. The right question targets what the company will be able to decide that it currently decides blind: where to cut cost, which customer to prioritize, when to hire.
3. Does the decision gain pay for the cost and time of implementation within a short window? Not “in theory, in three years.” A small business doesn’t have the cash to bet on an uncertain long-term return. If the system only pays off in an optimistic multi-year scenario, the problem it solves probably isn’t the company’s most urgent one right now.
If the answer to the three questions isn’t clear, the priority isn’t buying a system — it’s formalizing the most critical, simplest-to-design process first. Only after the process is defined does the system have something to run on top of.
Technology amplifies judgment, it doesn’t replace the lack of it
No system — management, customer service, or AI — creates discipline the company didn’t already have. It amplifies what already exists: if the process is good, the system makes it faster and more visible; if the process doesn’t exist, the system just makes the chaos more expensive and harder to undo later. The question that separates the two situations isn’t about the tool, it’s about the judgment of whoever is deciding to buy it.
Follow Eleva Tecnologia for more analysis on technology applied to business: follow @ElevaTechno on X or @elevatechnologies on Instagram, or learn about the group at elevatec.net/about.
